[jdev] Gaim and gnomemeeting using jabber

PUYDT Julien julien.puydt at laposte.net
Thu Dec 2 01:58:36 CST 2004

Le mercredi 01 décembre 2004 à 15:34 +0000, Richard Dobson a écrit :
> > Notice that voip-presence and im-presence are not entirely unrelated:
> No but its unrelated enough to not be correct as an extension to IM 
> presence, as I and others have repeatedly said.


> > I really think that:
> > * using presence isn't as stupid as you think ;
> Its not just me that thinks it, and as you are new to jabber I would suggest 
> you listen to the people that have been working with it and developing with 
> it for a very long time (I myself have been working with jabber stuff since 
> before there was even a JSF).

You must understand that people tend to be very very careful when said
"You really, really need to first design a protocol!" by a

> > * using jabber:x:oob to send the uri isn't against the standard ;
> No using jabber:x:oob is the correct thing to send the URI in, but you are 
> using jabber:x:oob in a way that is wrong and will potensially cause serious 
> problems with backwards compatibility, sending a voip-uri in jabber:x:oob is 
> asking the recieving client to execute that URI and pass it right to the 
> appropriate application on the users system, which will likely cause them to 
> start a voip call, trying to use this as a way to say you are in a voip call 
> is simply wrong and against the spec.

JEP-066, example 6?

> > * it works _now_, and not in 10 years when people will have finished
> > debatting the issue ;
> It wont work now due to the problems outlined above.

Well, I did test my patch, and it does work...

> > * it makes a very simple patch that has a chance to get upstream now ;
> It might be simple but it is still wrong.

JEP-066 doesn't call what I do 'wrong', it says 'creative usage' ;-)

> > Bad. You assume that deciding if the call is done or not is jabber's
> > problem: it's not. There's a voip-client there: it will ask the user for
> > confirmation. When someone has your voip-uri, you still have your word
> > to say.
> >
> > Besides, aren't the presence packets only sent to contacts for which I
> > already said they were authorized to know about me?
> Not always no, if for example you are in a chatroom the people in that room 
> will get your presence, and if you are adding your extension to presence 
> lots of people will know what your machines IP etc is, which is very bad, 
> jabber is designed in such a way that normally your IP address will not be 
> revealed to all and sundry, also when chatting to people you and adding them 
> to your contacts it doesnt mean you necessarily trust them enough to reveal 
> your ip address uncessarily, what about DDos attacks or the various buffer 
> overflow attacks that can be performed against a users machine once someone 
> has their IP? The VoIP software you speak of will do nothing to protect the 
> user from that.

Arg. Point taken.

> > Hmmm... as far as I remember, I had to allow my contacts to see me, and
> > they had to decide they wanted to see me... so telling them "I'm there"
> > isn't polluting the network, it's using it.
> But you are not just telling them you are there (thats what the standard IM 
> presence already does on its own), you are telling them extra information on 
> top of that which most will not want or need to know, thus the pollution, 
> its far better if only clients that want to receive said extra information 
> receive it, thus the need to use pubsub.

Hmmm... but if I tell them I have voip, and their client allows them to
get my voip-uri on demand... what's the difference?

> Rushing through things just because you are impatient is never a good idea, 
> start working with us rather than fighting us and you will find this will go 
> much faster.

Well, you'll notice that I wrote to ask, so I'm not that bad.

> There are several separate tasks you seem to be trying to bunch together 
> here and in doing so you are not doing it the right way, the different tasks 
> I can see are as follows:
> 1) advertise a client is voip capable (this should be done using JEP-0030 
> and JEP-0115).
> 2) initiate a chat with another voip capable client (this should be done 
> using JEP-0020, then JEP-0066).
> 3) advertise your voip-presence as something separate from the normal 
> im-presence (there is nothing currently to do this, but as has been said a 
> protocol based on pubsub is the best solution to this task).
> So overall you can do most of what you are trying to do now and right away 
> without having to wait for anything, it is only task 3 that needs a protocol 
> developed for it, and IMO step 3 is the least needed part of this and isnt 
> really even needed to get voip calling working in jabber clients, if I were 
> you I would just forget about task 3 for the moment and leave that for 
> sometime in the future, it really is not needed as you can use the 
> im-presence to notify other people you are on the phone (IMO you should be 
> just doing it this way anyway).

Step 3 can wait.

> Well if your "simple" patch is not following the standards I dont see them 
> even accepting that, if all you really want to do is allow people to call 
> each other using their im clients then you just need to implement Tasks 1 
> and 2 outlined above and just forget about Task 3 (which IMO is not needed 
> anyway).

Since the beginning of the thread, I modified the patch to send&accept
<x xmlns='jabber:x:oob'>
<url>h323:me at there</url>
in presence, and according to JEP-066, it follows the standard ; but
your objections are a concern :-/

Back to the chalkboard :-/


More information about the JDev mailing list